FR2664113A1 - METHOD OF DIGITAL TRANSMISSION WITH SPECTRUM SPREADING BY DIRECT SEQUENCE TO CHANGE SEQUENCES DURING TRANSMISSION, AND TRANSMITTER AND RECEIVER IMPLEMENTING IT. - Google Patents

Abstract

The receiver comprises: receiver and demodulator means (20), delivering from the received signal a succession of complex elements (s (kT)); means (26) for producing sequences of successive complex elements (Cj *) consisting of the conjugate elements of homologous elements used in transmission; means (27) for synchronizing the sequences in reception and in transmission; means for multiplying the complex elements (s (kT)) obtained at the demodulation output by a corresponding sequence of successive complex elements thus produced and synchronized; and summing means (28), for summing over the duration of an information symbol the resulting complex elements succeeding each other over this duration, so as to deliver an estimate of the initial information symbols used on transmission. According to the invention, it further comprises means (21, 22, 23, 34) for analyzing the complex elements (s (kT)) obtained at the demodulation output so as to determine a marking introduced on transmission, and selector means (25), controlled as a function of the marking thus determined so as to keep the same sequence of conjugated complex elements, or to change the sequence; the successive changes of sequences occur at random instants, determined by the receptor as a function of the sole analysis of said markings, but according to a deterministic order, known intrinsically by the receptor. The marking may in particular reside the value, identical or opposite, attributed to the first two elements of the sequence corresponding to each information symbol.

Description

Digital transmission method with spread spectrum by sequence

  The present invention relates to the protected transmissions of digital signals, in particular transmissions by

radio links.

  The purpose of protection is to make the connection resist

  aunt facing jammers; in this respect, one of the modes of protection

  possible well-known, to which the invention relates, is the

  said spread spectrum spread by direct sequence.

  This process is described, for example, in Jack's book

  K Holmes entitled Coherent Spread Spectrum Systems, a Wiley-

  Interscience Publication, John Wiley & Sons, 1982.

  Essentially, it consists in combining the symbols of

  training that must be transmitted to a pseudo-random sequence

  at a faster rate, known to the transmitter and the receiver when receiving

  the reverse operation is carried out so as to restore the sym-

  initial information bots used at issue or, at the very least, to provide an estimate of these symbols (taking into account

  noise and interference that has degraded the transmitted message).

  More precisely, the information symbols used are complex symbols corresponding to a quadrature amplitude modulation, and therefore comprising a real part and

  an imaginary part each consisting of one or more bits of

training.

  These complex symbols follow one another

  broken in the form of a flow symbol train Ds, as

illustrated figure 1.

  The direct sequence spread spectrum method consists of producing on transmission, together with the symbols

  information, (complex) elements Cp, called in the technical

  that "chips", of a pseudo-random sequence recurring The flow of chips is greater than the flow Ds of the symbols, the ratio

  Dp Ds = r being called "spreading gain" (for reasons of com-

  modified, we generally choose for r an integer value)

  complete sequence includes L chips, corresponding to a succes-

  assignment of L / r symbols Once the sequence of chips C j completed, it is immediately reiterated to be combined with the following symbols. The result of the combination between symbols ai and chips CQ will be a train of symbols if, at flow D chips, which will be used to

  modulate the communication transmitter.

  The passage from the higher rate D to the higher rate D produces, in the frequency domain, a spread of the spectrum of the transmitted signal, that is to say that the frequency band used is r times larger than the Nyquist band corresponding to the transmission

only information symbols.

  This results in: greater immunity to noise (in a ratio close to the spreading gain), excellent resistance to jammers, which will act only on a reduced fraction of the spectrum, given the spread of this spectrum. ci, and a lower detectability (for the same energy transmitted, it is distributed over a wider band and the spectral density of the emission is therefore

closer to that of noise).

  At the reception, the symbols if obtained at the exit of demo-

  In this case, the symbols, if received at the rate Dc, by the conjugates Ci * of the elements C of the sequence used at the time of transmission, also produced at the bit rate D, are multiplied for this purpose. then summing on r successive results; thus, at the rate Ds = Djr, an estimate of the initial information symbols ai is restored (in the absence of noise and jammers, we would reconstitute exactly

these symbols).

  It will of course be assumed that the synchronization between transmission and reception has been previously found, that is to say

  that we know how to recognize the moments of the beginning of the successive symbols

sive.

  This decoding will bring the enlarged band back to its original width.

  and conversely, will spread the jammer in a ratio r (the

  samples of the interfering signal will be multiplied by the

  elements of the sequence), which will therefore decrease

  its spectral density and thus its disruptive effect.

  This system provides excellent protection against jammers, but it nevertheless has the disadvantage that, if the same sequence is used permanently, the risks of interception and enemy intrusion are increased; this is especially the case if the sequence is generated in a simple way from a small number of elementary chips (thus, a PN sequence of length L = 1023 is generated from a key of only 10 bits): it will be then easy for an enemy to come to reconstruct this sequence and to

  intercept or corrupt the link.

  To mitigate this risk, it is generally

  ger of spreading sequence during transmission.

  Until now, in all the techniques proposed for this purpose, it was intended to warn the receiver of the change to come,

  either by interrupting the transmission of information and by

  a message to prevent a change of sequence, either in

  using a parallel channel, distinct from that used for the trans-

  mission of the information itself, in order to transmit to the

  receiver the sequence change information.

  These solutions are, however, cumbersome to implement

  (need for an additional channel or bidirectional link

  management of link breaks, etc.), which has hitherto

  limited their generalization to this type of transmission.

  One of the aims of the invention is to propose a method of spread spectrum transmission and sequence change allowing, with a simple unidirectional link,

  to transmit to the receiver, while transmitting, information

  possible changes in the spreading sequence, without degrading

  der nor interrupt this link at the moment of change.

  One of the aims of the invention is also to propose such a method that maintains a total transparency of the transmitted information, ie a method in which the information rate is not modified and in which it is not used, to signal the change, special symbols interposed between the information symbols themselves.

  For this purpose, the invention proposes to carry out a

  intrinsic (ie without addition of special symbols)

  the emitted sequence; the receiver will then detect the type of

  and make a decision based on the marking as well as

recognized.

  The absence of the use of special symbols has two advantages: first, it will not be necessary to filter these

  special symbols in the flow of information symbols

  received; this stream can therefore be used as is, without specific prior processing; second and most importantly, the sequence change

  can be detected at a very advanced stage of the process.

  sus (for example directly after demodulation), which will allow immediate action even before the decoding and processing of symbols. It will thus be possible, with a technically less complex device and a simple unidirectional link without any particular arrangement, to proceed to immediate sequence changes, occurring at random instants decided by the single transmitter and without anything in the connection (interruption, transmission of signals parallel, etc.) does not predict a change of sequence

  imminent for anyone trying to intercept the link.

  The security of the link will be increased in very large proportions, and with a very minimal equipment cost. For this purpose, the invention proposes a method of the aforementioned type with direct sequence spectrum spreading, ie in which:

  at the time of issue, at a given

  pseudo-random quantization of complex elements

  sifs, each of which is multiplied by the symbol

  information plex arriving concurrently at a lower rate, so as to obtain a succession

  resulting complex elements modulating a

  and on reception, we multiply the complex elements obtained at the demodulation output by a sequence

  successive complex elements consisting of

  conjugated elements of the homologous elements used on transmission, the transmission and reception sequences having been previously synchronized, and the duration of an information symbol is summed up by the resulting complex elements succeeding each other over this duration, so as to issue an estimate of symbols

  initial information used on the program.

  According to the invention, this process is characterized in that: - on emission, one selectively uses one or the other

  two sequences of neighboring complex elements,

  only by intrinsic marking, depending on whether the following sequence will be

  used will be the same as the sequence

  At the receiving end, the complex elements obtained at the demodulation output are analyzed so as to determine their marking, and it is decided to keep the same sequence of conjugated complex elements, or to change their sequence, as a function of the marking. that it has been recognized, the successive changes of sequences occurring at random instants, determined by the receiver according to the sole analysis of said markings, but in a deterministic order,

  known intrinsically by the receiver.

  Preferably, the marking lies in the value

  to a plurality of predetermined elements of the sequence, these values being chosen so as to maximize the distance between the

two types of marking.

  The predetermined elements of the sequence can in

  particular number of two for each information symbol.

  These two predetermined elements of the sequence can then be of opposite values in the first case of marking, and of

  same value in the second case of marking.

  To determine the marking, it is then sufficient to calculate the quadratic sum, over the duration of each information symbol,

  respective distances of the two predetermined predetermined elements

  during this period, and to compare this result with

a predetermined threshold.

  The invention also relates to a transmitter and a receiver.

  digital signal processor implementing this method.

  The invention will now be described in detail with reference to

reference to the appended figures.

  Figure 1, above, illustrates the way in which

  born the respective flows of symbols and chips.

  FIG. 2 is a block diagram of a transmitter transmitter

  both in the process of the invention.

  FIG. 3 is a block diagram of a metric receiver

  both in the process of the invention.

  In Figure 1, as indicated above, at each

  symbol ai corresponds r complex chips C 3 which will each be multi-

  folded by the (complex) value of the corresponding symbol present in

same moment.

  The marking according to the invention will consist in distinguishing two cases:

  the first case will correspond to an absence of change

  sequence, that is, the following sequence

  which will be used (after the Lth chip) will be the same as the current sequence; the second case will correspond to a change of sequence after the Lth chip of the current sequence, that is to say that the following sequence which will be used

  will not be the same as the current sequence.

  For this purpose, it is possible to operate on the value of certain well-determined chips corresponding to each of the symbols, for example

  the first two chips C 1, C 2 of the subsequence of r chips corre-

  nesting at the symbol ai, and likewise for all subsequent subsequences; in other words, we will operate on the value of all

  complex chips Cl + i r and C 2 + i r 'with i SE {0, 1 (L / r) -1).

  It should be noted, however, that the choice for the marking of

  the first two chips corresponding to each symbol is not limi-

  tive, and that one could as well choose two potato chips of some

  predetermined number, or a number of chips greater than two, etc. However, as will be explained below, the use of a number of chips greater than two may unnecessarily distort the

  sequence, and thus reduce the efficiency of information transmission.

  mation. For the marking, we can choose to give

  the second chip corresponding to each symbol an oppo-

  that is to say that we will have: C 2 + ir = Cl + i r in the first case mentioned above; and

  C 2 + i r = Cl + i r in the second aforementioned case.

  This choice has the advantage that, in "normal" mode,

  wrong "(no sequence change, which is statistically

  the most common case) is the sequence of the first case which will be used

  for spreading and, in its combined form, for the despreading

  the marking will not result in a decrease in the peak of

correlation at the reception.

  The only decrease in this peak will occur during the

  quence immediately preceding the change: indeed, at this moment, we will have used the sequence of the second case for the spreading but the conjugated sequence of the first case for despreading. However, if one uses only a small number of chips of marking compared to the number of chips per symbol (for example if one uses only two chips of marking for a gain of spreading r = 15 ), the decrease in the peak of correlation will be relatively low and the signal regeneration treatment will restore

  without difficulty the original symbol.

  The mode of marking chosen (one of the chips is either the opposite or the copy of another), although a priori non-limiting, and

  particularly advantageous in so far as it corresponds

  at the maximum Euclidean distance between the two cases.

  The use of maximum remote marking makes it possible to reduce to a minimum the rate of false decisions caused by noise and jammers. We also see that the sequence change information is very redundant, since it is repeated on all the symbols of the same sequence, ie L / r times, which gives it a

  very high resistance to noise and jammers (which also facilitates

  the synchronization of the reception detection).

  This number of occurrences of the sequence change information can be modified without departing from the scope of the invention;

  however, in practice, it is recognized that it is necessary

  r is at least 10 and the number of occurrences of

  the sequence change information exceeds 50.

  The transmitter, schematically illustrated in FIG. 2, is

  conventional construction, with a sequence generator 10 deli

  setting, among a set of stored sequences, a specific sequence

  mined, applied to a combinational circuit 11 performing the multiplication operation with the symbols to be encoded The output of this combinational circuit 11 constitutes the digital signal applied at the input of

emission modulator 12.

  Compared to a conventional transmitter with a change of

  sequence comprising a set of q sequences, the transmitter of the invention

  tion has a double number 2 q of sequences, namely q sequences carrying the aforementioned marking of the first case and q sequences, homologous to the previous ones, bearing the marking of the

second case.

  Before each sequence change, the transmitter

  will be a sequence marked according to the second case and then go to the

new sequence.

  If necessary, a "guard time" can be

  allowing the receiver time to take some

  actions before the sequence change.

  In this case, before changing the sequence, one will first use a sequence marked according to the second case (to prevent the

  receiver of the next sequence change), then one or more

  if similar sequences but marked according to the first case (corresponding to the guard time), then we will go to the new sequence. The sequencing of the sequences is deterministic, the sequences being sequenced according to a predefined order, for example circular looping, that is to say that after the sequence Sp is used.

  the sequence S Rp), with f Rp) = (p + 1) modulo q.

  On reception (FIG. 3), the successive sequential samples received at the output rate D of the demodulator will be examined. These sequential samples are of the form: ## EQU1 ## - (ir + j) T) + N (k T) + J (k T) ij = where T is the sampling period, where ai is the information symbol used on transmission, Ci.r + j being the current chip used for spreading, h (t) being the impulse response of the whole of the

  transmission chain (taking into account all the

  n (k T) being the sample resulting from the additive noise, and J (k T) being the sample resulting from the emission of the jammer. To recognize the marking, we will extract by means of

  of circuit 21 the samples corresponding to the only potato chips

  that is, when k takes the values (1 + i r) and (2 + i r).

  This treatment will be of a differential type in order to

  the impact of the change of symbol ai which occurs while

the r chips.

  The circuit 22 will then calculate the complex difference s (2 + i r) s (1 + i r), and the circuit 23 will then calculate the module

  It s (2 + i r) s (1 + i r) 11 of this complex difference.

  The accumulator 24 will then proceed to the quadratic summation of the modules thus calculated on all the symbols.

  corresponding to the same sequence.

  A variable is obtained: (L / r) -1 v = 1 l I s (2 + i r) s (1 + i r) 112 i = O which will be the decision variable of possible change

coding sequence.

  This variable can still be written: (L / r) -1 = I 1 s (m + 1) s (m) I 1 2 (with m = ir + 1) i = O (L / r) -i = ll ai (Cm + l-Cm) + (N (m + 1) N (m)) + (J (m + 1) J (m)) I 1 2 i = o (L / r) -1 = 1 ai (Cm + l-Cm) + (N (m + 1) -N (m)) + OJ (ej (m + 1) Tire Jm T)) 11 2 i = O 2 being the frequency of the jammer (it is assumed that one himself

  in the case of a pure carrier scrambler, and it is assumed that

  that the terms of intersymbol are null, that is to say that one

a: h ((m-n) T) = 3 minutes).

  This decision variable is not the only one possible, and the invention is not limited to this choice. In particular, if one wanted to privilege the noise resistance more than the resistance to the jammers, one could choose a variable of different decision,

more efficient.

  The mathematical expectation Elvl of this decision variable

  v may be increased as follows, No being the spectral power density of the Gaussian baseband noise: Elvl = l 1 / ((L Ir) -l) l, 11 s (m + 1) s (m) I 1 2 mc L (E l I l (Cm +, - Cm) I 2 l + 4 No + JO 2 sin (n T / 2)) The decision-making (sequence change or not) will be made by a circuit 25 performing the comparison of Elvl with

  a threshold corresponding to this increase.

  This circuit 25 will control the eventual change of sequence of the sequence generating circuit 26, which contains a set of q different sequences, which are identical to that contained in the generator 10 of the transmitter and which follow one another according to the

same deterministic law.

  In a manner known per se, the sequence delivered will be combined at 27 with samples s (k T) from the demodulator, the results of this combination being then subjected to operations

  conventional digital processing, regeneration, etc. in the cir-

cooked 28.

  It will be noted that, in a variant and without departing from the scope of the invention, instead of using for the decision criterion the difference

  modules of marked samples, one could use the different

  However, such a procedure is less effective in dealing with jammers. Alternatively, instead of taking the decision downstream of the accumulator module 24, that is to say until the marked samples corresponding to L / r successive symbols of the same sequence have been analyzed, it is possible to envisage make an intermediate decision before complete summation,

  that is to say, based only on a part of the samples

  (which are all redundant) of the sequence.

Claims (5)

  A method of transmitting digital signals consisting of a complex symbol train corresponding to a quadrature amplitude modulation, which method is a direct sequence spread spectrum type method, in which: on transmission, produced at a given rate (Dc) a   pseudo-random sequence of complex complex elements   cessives (C), each of which is multiplied by the complex information symbol (ai) arriving concurrently with   a lower flow rate (D), so as to obtain a success   transfer of complex elements resulting from (if) modulating a carrier, and upon reception, multiplying the complex elements (s (k T)) obtained at the demodulation output by a   sequence of successive complex elements (C *) constituting   the combined elements of the homologous elements used on transmission, the transmission and reception sequences having been previously synchronized, and the duration of an information symbol is summed up by the resulting complex elements succeeding each other on this   duration, so as to deliver an estimate of the   initial information boolets used on transmission, characterized in that: -on emission, one selectively uses one or the other   two sequences of neighboring complex elements,   only by intrinsic marking, depending on whether the following sequence will be   used will be the same as the sequence   and on reception, the complex elements (s (k T)) obtained at the demodulation output are analyzed   to determine the marking, and it is decided to   the same sequence of complex elements   or sequence change, according to the marking that has been recognized, the successive changes of sequences occurring at random instants, determined by the receiver according to the sole analysis of said markings, but in a deterministic order, known intrinsically by the receiver.   The process of claim 1, wherein the   is the value attributed to a plurality of elements pre-   determined from the sequence, these values being chosen so as to   maximize the distance between the two types of marking.   The method of claim 2, wherein said predetermined elements of the sequence are two in number. for each information symbol.   The method of claim 3, wherein the two predetermined elements of the sequence are of opposite values in the first case of marking, and of the same value in the second case. marking case.   The method of claim 3, wherein, for determining the marking, the quadratic sum, over the duration of each information symbol, is calculated from the respective distances of the two successive predetermined elements appearing during   this duration, and this result is compared to a predetermined threshold.   6 A digital signal transmitter using   process of one of claims 1 to 5, comprising:   means (10) for producing at a given bit rate (D) a pseudo-random sequence of successive complex elements (Cj),   means (11) for multiplying these elements   successive plexes by complex symbols of information   mation (ai) arriving concurrently at a lower rate (Ds), so as to obtain a succession of resultant complex elements (si), and modulator and emitter means (12), fed by the resultant complex elements thus obtained so as to producing a quadrature amplitude-modulated signal from these elements, which transmitter further comprises: selector means, for using one or the other of   two sequences of neighboring complex elements, different   only one of the other by an intrinsic marking, according to the following sequence which will be   used will be the same as the sequence   rante. 7 A digital signal receiver implementing   process of one of claims 1 to 5, comprising:   receiving and demodulating means (20), deli-   on the basis of the signal received, a succession of complex elements (s (k T)),   means (26) for generating sequence sequences   successive complex structures (C *) consisting of   conjugated elements of homologous elements used on transmission, means (27) for synchronizing the sequences thus used on reception on the homologous sequences used on transmission, means for multiplying the complex elements (s (k T)) obtained at the demodulation output by a   corresponding sequence of complex complex elements   cessives thus produced and synchronized, and summing means (28), to summon on the   duration of an information symbol the elements   resulting plexes succeeding over this period, so as to deliver an estimate of the initial information symbols used in the transmission, receiver characterized in that it further comprises: means (21, 22, 23, 24) for analyzing the complex elements (s (k T)) obtained at the demodulation output so as to determine a marking introduced on transmission, and selector means (25), controlled according to the marking thus determined so as to maintain the same sequence complex conjugate elements, or to change sequence, the successive changes of sequences occurring at random instants, determined by the receiver as a function of the only   analysis of the said markings, but in a   ministe, known intrinsically by the receiver.